[foaf-protocols] FOAF+TLS: RESTful Authentication for Distributed Social Networks

[Reto Bachmann-Gmür]

Story Henry said the following on 03/17/2009 03:26 PM:
...
> Secondly a name is a name of an object. But it is also impossible to
> separate the name from its sense and the sense from its referent,as
> Frege showed, and as much of 20th century analytic philosophy discussed.
> Names are handles on objects. They refer to the objects, but in a
> specific way. Otherwise if a person were known as Tim and as Timothy the
> following two statements would have exactly the same content
> 
> A: Tim = Tim .
> B: Tim = Timothy .
> 
> Still this is a complex question, and I don't think we can answer it in
> the paper given the space available. What is clear is that it makes
> sense to have a relation from a resource, identified via a URI to it's
> meaning (sense). 
I agree with you that this questions can't be answered in the paper, but
I went on thinking about the two statements and if they do not in fact
have exactly the same meaning.

In terms of RDF the two statements are clearly distinct as RDF semantics
know nothing about the semantics of owl:sameAs. In terms of owl
owl:sameAs expresses the identity relations which exists exclusively
between a thing and itself, so both statement express the necessary
truth (even though maybe not an a-prioric truth).

So I'm thinking if rather than an actual claim about the world "Tim =
Timothy" is a (imho a bit awkward) way to assign two names to a
resource, not sure if the mapping from URIs to their meaning is part of
the interpretation or if the difference between the two statements is
merely syntactic.

> I think RDF does also make a distinction between sense
> and referent. I have seen talk of this in a number of places, including
> owl, where one uses owl:equivalentClass in favor of owl:sameAs to
> distinguish the relation that relates two classes with the same elements
> from two classes with the same meaning.
It's easy to talk about two classes with the same extensions
(owl:equivalentClass), but two classes with the same meaning
(owl:sameAs) are not actually /two/ classes, they are one resource with
at least two names.

still going on reading the paper and wondering if this distinctions will
be relevant.

Reto